Insights from TAUS: AI, Machine Learning & Future-Proofed Translators
I was fortunate enough to be one of the 130 participants of the TAUS Global Content Conference in June of 2019, (and the only freelance translator). The two days that followed were full of networking, knowledge sharing, and inspirational speeches. TAUS Global Content Conference focuses on big questions such as how far machine translation (MT) has come, whether human parity is possible, how machine learning (ML) will impact global communication, and of course, what the rise of technology means for freelance translators.
What follows is my summary of the event and how what I learned can affect translators around the world. For more information, click here to read the footnotes for 2019’s TAUS Global Content Conference.
Lesson One: Neural Machine Translation (NML) Will Fundamentally Alter the Translation Process
Though we’ve had to deal with the rise of CAT and other early MT tools, most of these left the basic tenets of translation untouched. Jobs were still priced on a per-word basis and the translator was the only link between the source and final content. In short, the flowchart of a freelance translator’s tasks looked much the same as it did twenty or thirty years ago. Nowadays though, that’s no longer the case. Below are a few ways that upcoming technologies will permanently alter the translation process:
- Pricing on a per word basis is no longer practical: Instead the rates and payments may work better using hourly fees or project based fees.
- You’re more likely to serve as a proofreader than a translator for some jobs: While people might not be willing to pay a high-end translator to do the grunt work, they’re still open to bringing you on for an editing role.
- Your next project manager might be an algorithm: With automatically generated content where deadlines and algorithms drive decisions, the need for dedicated project managers will shrink. The data generated from algorithms will then be used to drive recruitment, predict market readiness, and determine translation needs. In other words, a computer could one day be your boss.
- Multimedia translation is the future: With VR and AR on the rise, the need for translators willing to adapt to these new technologies will increase. Translators who are willing to embrace these different platforms, including voice-assisted search, will benefit from our early adaptation.
Lesson Two: The Proper Application of MT and AI in Your Workflow Can Set You Apart from Your Competition
As language specialists, we tend to think of MT and AI as the monster under our bed. But if we are willing to embrace these monsters we can use them to our advantages. Having a good MT on hand can help cut costs and drastically increase our production ability, at least if we know how and when to use it.
Lesson Three: Find Your Niche to Future Proof Your Business
John Tinsley of Iconic Translation Machines opened with a startling statistic: ”99% of today’s global content is translated by MT without any human intervention.“ Hearing numbers like that is enough to leave most translators waking up in a cold sweat. I’m not one of them. My research shows that there’s plenty of room for both machine and human translators in the marketplace. In fact, experts estimate the employment outlook for translators and interpreters will increase 29% by 2024. Of course many of you are probably wondering what you can do to ensure you’re positioned for growth.
Generalists and low-quality content mills will be the main casualties in this battle of man versus machine. To succeed in the future, you’ll need to specialize in a field that values accuracy and creativity over speed and volume. Examples of these fields include medical oncology research, renewable energy development, and automotive patent law as well as more creative fields such as marketing. Transcreators, because of their ability to merge copywriting with nuance and emotion, will also be immune to the rising tide of automation.
There is also a new field of specialization to pay attention to. International companies are hungry for freelance translators with experience in digital marketing and social media. They need people who understand local search and can adjust their messaging to meet the needs of customers and search engines alike. Even if an oncology degree is out of reach, the internet is awash with knowledge about customer service, search engine optimization, and social media management.
Lesson Four: There’s a New Translation Business Model on the Rise
Up until a few years ago, translators came in two main stripes: those who collaborated directly with clients and those who worked for language service providers. Many freelance translators make a living straddling the line between the two. These two different ways of working demand different technologies, processes, and timelines. Now, many translators will be forced to contend with a third translation model: crowdsourcing.
In this model, popularized by platforms such as Smartcat and Gengo, end-clients plug their content into a virtual translator community. The platform then automates job management on both sides. By removing the geographical barriers between translator and end-client, the crowdsourcing model opens the market to lower-cost providers in markets where the cost of living is much lower. These lower costs will ultimately hurt the language service providers stuck in the middle. This new business model gives you even more reason to specialize and find your niche.
Lesson Five: Machine Learning Is Creating New Roles for Translators
If generalists and low-quality translators will lose their place in the market, new, more specialized people will take their place. In the next few years, we expect to see a sharp increase in demand for:
Post Editors of MT
While some companies don’t mind machines doing the lion’s share, they’re skittish about releasing raw machine translations into the market. Therefore, they’ll want someone to read over the outputted text for mistakes. The corrections will then be fed back into the data pipeline to further improve the machine’s output. This is a possible field for translators willing to take it on, which can even create a new type of translator.
Transcreators
Machines still struggle to combine culture, language, and emotion into a creative and compelling text. To get that done right, you need a human. Machines also have a tough time understanding cultural nuances, emotions, and the psychology involved in proper translation. All things that are vital to sales and marketing. For that reason, transcreators will weather the digital transformation just fine.
Spoken Content Specialists
Audio is being hyped as the new communication platform. Over time, text interfaces are being replaced with voice-powered assistants like Google Home and Alexa. Podcasts are also growing in popularity. This shift to audio requires a different sort of translator. Those who work in this field must find ways to convey the importance of social, gender, and emotive markers in the work that they produce. They must also be prepared to work with speakers of various dialects in order to create more localized, regionalized content.
Brand Ambassadors
With more of the world coming online, globalized branding is becoming more important. Worldwide retailers will need help ensuring that their company’s values and beliefs are effectively communicated across geographical borders. This will require translators who are willing to act as product testers and advise companies on how best to adjust their product to meet the needs of their target markets.
Local Storytellers
Good marketing is built around good narratives. Simply translating these stories in a target language rarely has the desired impact that companies want. That’s why these groups are looking for people who know how to craft a good tale. This is a perfect role for the translators who also dreamt of a career as a creative writer. And, with the demand for personalization growing, the need for this type of work will only grow.
Conversational Agent Consultants
Like it or not, chatbots are here to stay. While much of their development has happened in English-speaking countries, they’re now being deployed in markets all around the world. Still, these “conversational agents” need a human to help fine-tune and perfect them. People in this position will have to have a good understanding of human psychology, machine logic, and customer service to succeed.
Lesson Six: Machine Learning Will Give You Insight into Your Work
Translators, as connected professionals, will eventually benefit from the datafication of their work. The rise of algorithm and machine translation will make it easier for us to analyze our work habits, performance levels, and productivity. Eventually, ML could allow us to build a data narrative of our past choices and spell check preferences. It could then use that data to make suggestions to improve the overall quality of our work. With this data at our fingertips, we’ll also have a brand-new way to quantify our value to clients and potential employers.
Lesson Seven: Traditional Training Isn’t Enough
Academic training for translators usually consists of two things: translation theory and practical translation. The latter tends to focus on computer-assisted translation tools and general customer service. The former focuses more on ethics, linguistics, and sociology. The topic of MT is barely more than a footnote in these areas. The curriculum rarely makes room for the advanced algorithmic knowledge, specialized skill sets, and advanced IT tools required for the future. Until more universities adapt to this, new translators will have to turn to industry mentors and less traditional sources to help prepare them for the future of the industry.
Lesson Eight: We Need to Stop Being Afraid of MT and Step Up to Help Shape It
Machine translation tools are often built by software engineers with little to no knowledge of the industry itself. This often results in tools that are clunky, complicated, and poorly designed for the translation process. But just like we did with bilingual print dictionaries and translation memories, we have a chance to change that. We can reach out to designers and engineers and ask for things to change. Or perhaps we can join a group or become guinea pigs for the next-generation tools coming down the pipeline. Whatever you do, don’t let the advent of ML leave you voiceless. Do everything you can to make sure that the end product you use is well suited to your work.
Lesson Nine: The Need for Globalized Content Has Never Been Greater
With ML on the rise, most of us think that the need for skilled translators will decrease. That couldn’t be farther from the truth. As huge population centers become increasingly digitized, there’s now an insatiable need for translators willing to work in those markets. Furthermore, with all that big data flying around, there’s a huge need for people to translate and draw conclusions across cultural lines. In short, while machines will take away a lot of the grunt work, freelance translators will be given more opportunities to make a difference.
To summarize: translators need to up their game, focus on niche markets, use MT as a way to boost productivity, use MT as a way to provide ancillary services, and also consider adjusting their business models or pricing schemes to keep MT profitable.
Attending TAUS was an eye-opening experience, being the “linguist” among technology companies. I not only learned a lot about the present state of the industry but also gained a lot of insight into what’s coming. If we specialize and use our human powers to our advantage, we have no reason to be scared of MT and its associated impacts.
Check also:
Swedish 
English